Coal surface mining in northern Great Plains USA led to reclamation experiments with soil respreading. Respread soil depth (RSD) and runoff of water redistribution (WR) effects interacted in original North Dakota studies, complicating interpretations. We determined WR and soil depth/soil quality (SQ) effects on hillslope production patterns for sites with soil wedges (2%–5% slope, 50-m length) over sodic mine spoils. At Zap, cool-season forages crested wheatgrass (CWG:
A need for land reclamation was created by the advent of coal surface mining in the northern Great Plains of the USA. In portions of this region Paleocene strata can give rise to mine spoil high in exchangeable sodium and clay content making revegetation difficult [
In one experiment near Stanton ND [
(a) One replication out of three of reconstructed soils at Stanton site. Overall length of construct was 243 m and average slope was 4%. The 3 : 1 mixed subsoil : topsoil treatment is not discussed in this paper. (b) Reconstructed soil at Zap site. Modified from Doll et al. [
In another experiment near Zap ND [
The influence of soil depth where water redistribution influence should have been minimal was shown by results of Barth and Martin [
Runoff of water to lower areas of hillslopes results in higher yields downslope.
Working on pastures dominated by cool-season grasses in Iowa USA, Harmoney et al. [
Results of North Dakota-based reclamation experiments, particularly in terms of production responses to respread soil depth over mine spoils [
The basic definition of SQ is the capacity of a soil to function [
Indices that quantify characteristics of landforms have been reviewed [
The objective of this study was to determine the relative effects of SQ measures (soil depth and SQI) and apparent water redistribution (represented by WRI) upon hillslope plant production patterns on reclaimed mine soils. Other factors whose effects on plant production patterns have been studied here were plant species and structure of reclaimed soils.
Experimental reclamation sites were located in Mercer county, North Dakota, USA, near the towns of Zap and Stanton at 47°14′33′′N-101°51′1′′W and 47°15′30′′N-101°20′55′′W, respectively. Mean annual precipitation was 421 mm, with 270 mm occurring during the 6-mo April–September growing season, and the mean 6-mo growing season temperature was 15.6°C.
Coal surface mining at the sites produced spoil from Mesozoic strata consisting of loosely consolidated siltstone or shale material that is saline and sodic. At the Zap site, mine spoil was leveled in 1974-1975 to a 2% north-facing slope. After removal and stockpiling of nearby Haplustoll topsoil material (mostly A-horizon), three types of subsoil materials (mostly B- and C-horizon) were placed in a double wedge-shaped hillslope with 21-m wide main treatment (subsoil quality) areas with two replications (Figure
Soil properties at experimental reclamation sites. AWC, available water content; EC, electrical conductivity; SAR, sodium adsorption ratio; SOC, soil organic carbon.
Zap reclamation site | Stanton reclamation site | |||||||
---|---|---|---|---|---|---|---|---|
Topsoil | Subsoil A | Subsoil B | Subsoil C | Minespoil | Topsoil | Subsoil | Minespoil | |
Sand, g kg−1 | 410 | 82 | 315 | 679 | 80 | 526 | 246 | 74 |
Silt, g kg−1 | 349 | 443 | 420 | 187 | 482 | 341 | 425 | 563 |
Clay, g kg−1 | 241 | 475 | 265 | 134 | 438 | 133 | 329 | 363 |
Texture, USDA system | Loam | Silty clay | Clay loam | Sandy loam | Silty clay | Sandy loam | Clay loam | Silty clay loam |
AWC, kg kg−1 | 0.174 | 0.147 | 0.149 | 0.096 | 0.128 | 0.190 | 0.157 | 0.157 |
Bulk density, g cm−3 | 1.211 | 1.251 | 1.331 | 1.448 | 1.353 | 1.279 | 1.279 | 1.340 |
| 1.32 | 4.76 | 2.29 | 0.79 | 5.86 | 0.60 | 2.99 | 3.05 |
SAR, (mol m−3)1/2 | 4.86 | 7.67 | 6.77 | 5.46 | 17.30 | 3.47 | 9.96 | 16.40 |
Saturation% | 41.2 | 60.1 | 48.5 | 27.4 | 99.6 | 41.0 | 62.7 | 112.4 |
| 7.16 | 7.36 | 7.73 | 7.07 | 7.54 | 7.37 | 7.91 | 7.95 |
SOC, g kg−1 | 14.3 | 9.7 | 7.1 | 7.2 | 8.5 | 15.6 | 8.1 | 9.5 |
Total | 1.16 | 0.80 | 0.60 | 0.69 | 0.71 | 1.3 | 0.7 | 0.3 |
available P, mg kg−1 | 9.2 | 11.4 | 3.6 | 6.7 | 7.1 | 8.7 | 4.0 | 4.9 |
Topsoil (mostly A-horizon) was spread over subsoils to a depth of 20 cm. At the Stanton site, spoil was leveled in 1974 to a 0.5% slope and a single wedge of subsoil material was formed. Main treatment strips 20-m wide with three replications were created by placement of 0-, 0.2-, and 0.6-m thicknesses of topsoil over subsoil (Figure
Each soil type block was divided lengthwise and four crop treatments were seeded in a randomized pattern within replications. Species seeded at Zap in 1976 were the perennial, cool-season grasses Russian wildrye (RWR;
Crops were harvested from 1976 to 1979 at the Stanton site and 1976 to 1981 at the Zap site, but, for purposes of this paper, data from 1978 and 1979 are used because they were years without drought in which both sites were represented. Vegetation was harvested from areas 4.3 m long by 1.8 m wide at 10 or 11 positions along each aspect of hillslopes, and dry weights of forage species’ aboveground herbiage and spring wheat seed yields were measured. Forage grasses were harvested in late June or early July, and ALF was harvested twice, in late June or early July and again in August or early September. Spring wheat seed was harvested from the later part of July to the first part of August.
Relative yields were calculated for each seeding treatment based on the yearly average for the treatment at a site. Yield patterns of individual forage species’ treatments were compared using running averages of three contiguous hillslope positions (1, 2, 3; 2, 3, 4; etc.). Applying one-way ANOVA to such averages normalized to annual treatment means, only 6 out of 132 site-position-soil treatment-year combinations indicated significant yield differences. Therefore, relative yields for a given year and site were aggregated among forage species treatments for further analyses.
Soil samples were collected by hydraulic probe in depth increments of 0–15, 15–30, 30–60, 60–90, and 90–120 cm. Samples for electrical conductivity (EC) and pH were collected at four positions from toe to summit along each hillslope aspect and those for SOC and available P at either one or two positions.
Soil chemical measurements were conducted by methods detailed in Black et al. [
Details about samples used for textural analysis are given in Wick [
Soil quality assessment was carried out by calculation of a SQ index (SQI) using the SMAF tool developed by Andrews et al. [
In order to calculate a SQI value for each of the 8 or 10 positions along hillslopes at which plant yields were measured, soil property values were estimated from measured values. For the chemical properties EC, SAR, and pH, measured at four locations, this was done by interpolation on treatment averages for each sampling depth increment. For the properties SOC, available P, and AWC, treatment averages for reconstructed soil horizons (topsoil, subsoil, or mine spoil) and soil depths were used to derive suitable values.
Water content measurements were examined for indications of downslope water redistribution. Neutron moisture meters had been used in steel access tubes, which had been installed two per subplot at a lower midslope and a near-summit position. Although measurements were made over multiple years, only 1979 data had been analyzed by individual depth increments in predecessor studies, making them suitable for indicating downslope water redistribution. Out of measurements made five times during the year, only those in late April-early May showed substantial evidence of water redistribution in the 0–30 cm soil depth. (The 0–30 cm depth was the only increment where water content differences between the upper and lower hillslope positions had not been affected by systematic variation in respread soil depth.) ANOVA was done across soil treatments within seeding treatments. Averaged across soil treatments, 0–30 cm soil water at Zap under CWG was 14% greater at the lower hillslope position (water content = 0.21 m3 m−3,
Moore et al. [
One-way ANOVA of individual seeding treatments’ yields used SPC for Excel (BPI Consulting; Cypress, Texas, USA). Soil water data was analyzed within site, depth, and seeding treatment by ANOVA with hillslope location as split plot within soil treatment using SAS 79 (SAS Enterprises Inc., Cary, North Carolina, USA). Correlations among soil properties and indices and regressions of plant production versus WRI and SQ metrics used SPC for Excel. Regressions of plant production versus hillslope position used Sigma Plot (Systat Software, Inc., San Jose, California, USA).
Growing season precipitation at Zap and Stanton during the years of the study was average or above average (Table
Precipitation at experimental reclamation sites near Zap and Stanton, North Dakota, USA. Annual median precipitation near the sites at Beulah ND for 29-year period was 39.2 cm.
Month | Zap | Stanton | |||
---|---|---|---|---|---|
1978 | 1979 | 29-yr median | 1978 | 1979 | |
cm | |||||
April | 4.4 | 4.5 | 3.4 | 4.2 | 3.2 |
May | 5.8 | 2.0 | 5.7 | 5.2 | 2.7 |
June | 6.4 | 6.1 | 7.5 | 8.8 | 7.2 |
July | 7.8 | 11.4 | 6.1 | 6.2 | 7.9 |
August | 1.5 | 2.3 | 3.0 | 0.9 | 12.3 |
September | 4.1 | 3.2 | 3.2 | 8.8 | 2.5 |
April through September | 29.9 | 29.5 | 28.9 | 34.1 | 35.8 |
Except for ALF, yields of perennial forages and spring wheat at both sites were considerably higher in 1978 than in 1979, reflecting a better distribution of spring precipitation during 1978 (Table
Biomass yields of perennial forage species or seed yield of spring wheat.
Site | Year | Crested wheatgrass | Native grass mix | Russian wildrye | Alfalfa | Spring wheat |
---|---|---|---|---|---|---|
kg ha−1 | ||||||
Zap | 1978 | 3103 | 2134 | 811 | ||
1979 | 1418 | 1149 | ||||
| ||||||
Stanton | 1978 | 4023 | 1462 | 2194 | 1639 | |
1979 | 1599 | 861 | 2327 |
Three out of six 1978-79 regressions of aggregated CWG and RWR yields versus hillslope position on the Zap 5% north-facing slope had about 20% to 45% increases with greater soil depths and elevations to 10 m or more from the toe of the soil wedge (Figure
Relative yields of perennial forage crops versus hillslope positions for soil treatments at Zap site during 1978 and 1979 showing means, standard errors, and cubic regressions. Yields of crested wheatgrass
On the 4% south-facing slope at Stanton, 1978-79 relative yields of the three (ALF, CWG, and NAT) forage species aggregated showed up to about 60% increases of production with increasing soil depth and elevation to midslope and then smaller decreases of yield further upslope (Figure
Relative yields of perennial forage crops versus hillslope positions for soil treatments at Stanton site during 1978 and 1979 showing means, standard errors, and cubic regressions. Yields of alfalfa
Hillslope patterns of perennial forages’ yields have been aggregated across soil treatments and compared with those of spring wheat seed yields for the year 1978 in Figure
Relative yields versus hillslope positions comparing perennial forage crops’ biomass and spring wheat seed for 1978 at Zap and Stanton sites showing means, standard errors, and cubic regressions. Soil treatments have been aggregated.
Hillslope water redistribution generated by runoff is sensitive to slope [
To understand results of regressions of plant productivity with measures of SQ and water redistribution, we examined correlations among various soil properties and indices (Table
(a) Correlations between distance from bottom (toe) of hillslopes and water redistribution index (WRI), respread soil depth (RSD), and soil quality index (SQI). (b) Correlation matrices for WRI, RSD, and SQI. (c) Correlations between SQI and soil properties. All
Site | WRI | RSD | SQI | |
---|---|---|---|---|
Zap | Meters from south toe | −0.925 | 0.979 | 0.846 |
Zap | Meters from north toe | −0.978 | 0.985 | 0.934 |
Stanton | Meters from south toe | −0.981 | 0.930 | 0.458 |
Zap, | WRI | RSD | SQI | Stanton, | WRI | RSD | SQI |
---|---|---|---|---|---|---|---|
WRI | 1 | −0.762 | −0.704 | WRI | 1 | −0.903 | −0.444 |
RSD | - - - | 1 | 0.890 | RSD | - - | 1 | 0.689 |
SQI | - - | - - | 1 | SQI | - - | - - | 1 |
site | SAR | EC | pH | AWC | SOC | P | |
---|---|---|---|---|---|---|---|
Zap | SQI | −0.904 | −0.734 | −0.464 | ns | ns | 0.221 |
Stanton | SQI | −0.861 | −0.915 | −0.859 | 0.719 | 0.579 | 0.376 |
At Zap, subsoil B and C treatments had positive regressions of forage species production with water redistribution index (WRI); the subsoil A regression was near significant (Table
Linear and multilinear regressions of perennial forages’ relative yields in 1978-79 versus water redistribution index (WRI), respread soil depth (RSD), and soil quality index (SQI) for soil treatments at Zap and Stanton sites. The third column in each group indicates the sign of regression coefficient(s). For entries with
| | | Coefficient sign | | | Coefficient sign | | | Coefficient sign |
---|---|---|---|---|---|---|---|---|---|
Zap 1978-79 | Zap 1978-79 | Zap 1978-79 | |||||||
Subsoil A | Subsoil B | Subsoil C | |||||||
| |||||||||
WRI | 0.103 | 0.123 | 0.003 | P | 0.217 | <0.001 | P | ||
| |||||||||
RSD | 0.074 | 0.021 | N | 0.197 | <0.001 | N | 0.188 | <0.001 | N |
RSD + WRI | 0.074 | 0.069 | N; ns | 0.197 | <0.001 | N; ns | 0.225 | <0.001 | ns; P |
| |||||||||
SQI | 0.075 | 0.020 | N | 0.187 | <0.001 | N | 0.221 | <0.001 | N |
SQI + WRI | 0.076 | 0.066 | N: ns | 0.195 | <0.001 | N; ns | 0.258 | <0.001 | N; P |
| |||||||||
Stanton 1978-79 | Stanton 1978-79 | Stanton 1978-79 | |||||||
0 cm topsoil | 20 cm topsoil | 60 cm topsoil | |||||||
| |||||||||
WRI | 0.830 | 0.109 | 0.583 | ||||||
| |||||||||
RSD | 0.418 | 0.479 | 0.639 | ||||||
RSD + WRI | 0.113 | 0.250 | <0.001 | P; P | 0.303 | <0.001 | P; P | ||
| |||||||||
SQI | 0.386 | 0.243 | 0.284 | <0.001 | P | ||||
SQI + WRI | 0.103 | 0.418 | <0.001 | P; P | 0.402 | <0.001 | P; P |
Multilinear regressions combining RSD or SQI with WRI also had negative coefficients for the SQ measure part of the equations. Positive regressions with WRI and negative coefficients for regressions with SQ measures reflected evident dominance of water redistribution effects on hillslope yield patterns.
In contrast to Zap, only multilinear regressions of yield versus RSD and WRI or yield versus SQI and WRI were significant at Stanton and then only for the two treatments with topsoil (Table
Regressions of yield versus WRI and SQ measures comparing 1978 spring wheat seed and perennial forage biomass indicated that spring wheat at Zap displayed a pattern of results similar to those of both spring wheat and perennials at Stanton (regressions aggregated across soil treatments; Table
Linear and multilinear regressions of 1978 spring wheat and perennial forages’ relative yields versus water redistribution index (WRI), respread soil depth (RSD), and soil quality index (SQI) aggregated across soil treatments at Zap and Stanton sites. The third column in each group indicates the sign of regression coefficient(s). For entries with
| Zap 1978 all soils | Zap 1978 all soils | Stanton 1978 all soils | Stanton 1978 all soils | ||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
| | | | |||||||||
| | coefficient sign | | | coefficient sign | | | coefficient sign | | | coefficient sign | |
WRI | 0.349 | 0.186 | <0.001 | P† | 0.057 | 0.021 | N | 0.492 | ||||
| ||||||||||||
RSD | 0.081 | 0.003 | P | 0.200 | <0.001 | N | 0.204 | <0.001 | P | 0.092 | 0.004 | P |
RSD + WRI | 0.119 | 0.001 | P; P | 0.220 | <0.001 | N; ns | 0.351 | <0.001 | P; P | 0.311 | <0.001 | P; P |
| ||||||||||||
SQI | 0.047 | 0.024 | P | 0.257 | <0.001 | N | 0.285 | <0.001 | P | 0.314 | <0.001 | P |
SQI + WRI | 0.055 | 0.052 | P; ns | 0.268 | <0.001 | N; ns | 0.285 | <0.001 | P; ns | 0.352 | <0.001 | P; P |
Differences in hillslope production patterns and regression results between forage species and spring wheat at Zap appear to be related to when the species are most actively using soil water and growing during the spring and summer. Perennial grasses at Zap, CWG, and RWR are both cool-season species, actively growing in earlier spring. Under North Dakota USA conditions, CWG reaches boot stage by late May to about June 1 [
Two out of three forage species at Stanton were either warm-season (NAT mix) or full season (twice-cut ALF). As noted above, there were very few differences among hillslope growth patterns of individual forage species, and this applied to cool-season CWG compared to the other forage species at Stanton. Thus, the principal reason for differences in hillslope growth patterns between perennials at Zap on one hand and perennials at Stanton plus spring wheat at both sites appears to involve seasonality of growth and water use of the plant species at the sites.
Aspect is another factor that could have been involved in the differences in hillslope yield patterns between the sites. The Stanton site had a single 4% south-facing slope, making it less responsive to earlier-season water redistribution effects on plant growth than the 5% north-facing slope at Zap. Examples exist in literature showing significantly greater soil water at north aspect positions compared to south aspect on Great Plains USA agricultural lands [
Current SQ assessment practice emphasizes management effects by examination of the most dynamic part of the profile near the soil surface, with emphasis placed on biological properties [
Especially in subhumid or semiarid areas, use of soil water deeper in the rootzone can become important under limited precipitation. Evidence for the benefit of considering the whole profile in SQ assessment was given in Merrill et al. [
Much of the disturbed land in the northern Great Plains region, such as that used for research here, has been or will be reclaimed to rangeland/pastureland usage. According to Herrick et al. [
Our results have shown that interpretation of reclamation experiments by soil salvage and respreading requires consideration of both soil depth through SQ analysis and basic hillslope water redistribution process. Analysis of hillslope growth patterns and regression analyses have shown qualitative differences between the sites and among plant species types. These differences may be summarized: more dominance of apparent water redistribution effects for cool-season grasses at Zap site and dominance or more balanced influence of soil depth and associated SQ factors for spring wheat at both sites and forage species at Stanton.
To obtain more quantitative analyses of reclamation experiment results such as those studied here, we suggest an analysis model be implemented that combines hydrologic processes on the hillslope scale with plant growth responses to soil factors at the point scale. For example, hydrologists have described and implemented a distributed hydrological model [
We studied mined land reclamation experiments in the state of North Dakota, USA, in which salvaged topsoil and subsoil materials were respread on high-sodium mine spoils, forming low hillslopes. At the Zap site, hillslope yield patterns of cool-season forage grasses CWG and RWR were more dominated by apparent water redistribution effects (yield increasing downslope) than by growth responses to soil depth increasing upslope. Hillslope yield patterns of forage crops (ALF, CWG, and NAT mix) at Stanton site and of spring wheat at both sites were more dominated by response to increasing soil depths from toeslope to midslope areas, with lesser response to apparent water redistribution from midslope to shoulder slope areas.
A soil quality index (SQI) based on six soil indicator properties was highly correlated (
The dominance of water redistribution effects on hillslope yield patterns at Zap, supported by regression analyses, appeared to be due to response of cool-season forage species to springtime runoff. Forages at Stanton and spring wheat at both sites are known to use soil water later in the season, and this was associated with hillslope yield patterns that were dominated by soil depth and SQ effects.
Our results show that assessment models for disturbed lands reclamation by soil respreading need to include information about plant community composition and phenology of growth and water use along with landform hydrology and SQ-soil structure information.
Inclusion of branded product information is for the benefit of the reader and does not imply preference nor endorsement by USDA-ARS.
The authors declare that there are no conflicts of interest regarding the publication of this paper.
The authors wish to acknowledge the leadership of the late Dr. J. F. Power in the mined land reclamation experiments upon which this paper is based.